Canada

The Vision Transformer (ViT) excels in accuracy when handling high-resolution images, yet it confronts the challenge of significant spatial redundancy, leading to increased computational and memory requirements. To address this, we present the Localization and Focus Vision Transformer (LF-ViT). This model operates by strategically curtailing computational demands without impinging on performance. In the Localization phase, a reduced-resolution image is processed; if a definitive prediction remains elusive, our pioneering Neighborhood Global Class Attention (NGCA) mechanism is triggered, effectively identifying and spotlighting class-discriminative regions based on initial findings. Subsequently, in the Focus phase, this designated region is used from the original image to enhance recognition. Uniquely, LF-ViT employs consistent parameters across both phases, ensuring seamless end-to-end optimization. Our empirical tests affirm LF-ViT&#39;s prowess: it remarkably decreases Deit-S&#39;s FLOPs by 63% and concurrently amplifies throughput twofold. The supplementary source code is made available for in-depth examination.

AAAI 2024

LF-ViT: Reducing Spatial Redundancy in Vision Transformer for Efficient Image Recognition

efficient computing

vision transformer (vit)

adaptive inference

The Vision Transformer (ViT) excels in accuracy when handling high-resolution images, yet it confronts the challenge of significant spatial redundancy, leading to increased computational and memory requirements. To address this, we present the Localization and Focus Vision Transformer (LF-ViT). This model operates by strategically curtailing computational demands without impinging on performance. In the Localization phase, a reduced-resolution image is processed; if a definitive prediction remains elusive, our pioneering Neighborhood Global Class Attention (NGCA) mechanism is triggered, effectively identifying and spotlighting class-discriminative regions based on initial findings. Subsequently, in the Focus phase, this designated region is used from the original image to enhance recognition. Uniquely, LF-ViT employs consistent parameters across both phases, ensuring seamless end-to-end optimization. Our empirical tests affirm LF-ViT's prowess: it remarkably decreases Deit-S's FLOPs by 63% and concurrently amplifies throughput twofold. The supplementary source code is made available for in-depth examination.

technical paper

We are pleased to announce the Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI-24), which will be held in Vancouver, British Columbia at the Vancouver Convention Centre – West Building from 20-27 February, 2024.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-24 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.

We expect for AAAI-24 to be an in-person conference – one author of all accepted papers will be expected to present work in person unless there are exceptional circumstances that prevent this.<br><br><br><br>

In order to access the AAAI-24 event page you need to register [here](https://aaai.org/aaai-conference/registration/)

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. 

This paper tackles the problem of federated domain generalization in person re-identification (FedDG re-ID), aiming to learn a model generalizable to unseen domains with decentralized source domains. Previous methods mainly focus on preventing local overfitting. However, the direction of diversifying local data through stylization for model training is largely overlooked. This direction is popular in domain generalization but will encounter two issues under federated scenario: (1) Most stylization methods require the centralization of multiple domains to generate novel styles but this is not applicable under decentralized constraint. (2) The authenticity of generated data cannot be ensured especially given limited local data, which may impair the model optimization. To solve these two problems, we propose the Diversity-Authenticity Co-constrained Stylization (DACS), which can generate diverse and authentic data for learning robust local model. Specifically, we deploy a style transformation model on each domain to generate novel data with two constraints: (1) A diversity constraint is designed to increase data diversity, which enlarges the Wasserstein distance between the original and transformed data; (2) An authenticity constraint is proposed to ensure data authenticity, which enforces the transformed data to be easily/hardly recognized by the local-side global/local model. Extensive experiments demonstrate the effectiveness of the proposed DACS and show that DACS achieves state-of-the-art performance for FedDG re-ID.

Diversity-Authenticity Co-constrained Stylization for Federated Domain Generalization in Person Re-identification | VIDEO

Spatiotemporal predictive learning is a learning paradigm that enables models to learn spatial and temporal patterns by predicting future frames from given past frames in an unsupervised manner. Standard methods rely on recurrent units to capture long-term temporal dependencies but often require heavy computational costs and underperform in real-world scenarios. This paper presents an innovative Wavelet-based SpatioTemporal (WaST) framework for predictive learning. We extract and adaptively control both low and high-frequency components at image and feature levels via 3D discrete wavelet transform for faster processing while maintaining high-quality predictions. Consequently, we introduce a time-frequency aware translator uniquely crafted to efficiently learn short- and long-range spatiotemporal information by individually modeling spatial frequency and temporal variations. Meanwhile, we design a wavelet-domain High-Frequency Focal Loss that effectively supervises high-frequency variations in the wavelet domain. Extensive experiments across various real-world scenarios, such as driving scene prediction, traffic flow prediction, human motion capture, and weather forecasting, demonstrate that our proposed WaST achieves state-of-the-art performance over various spatiotemporal prediction methods.

Wavelet-Driven Spatiotemporal Predictive Learning: Bridging Frequency and Time Variations

Generating photorealistic 3D faces from given conditions is a challenging task. Existing methods often rely on time-consuming one-by-one optimization approaches, which are not efficient for modeling the same distribution content, e.g., faces. Additionally, an ideal controllable 3D face generation model should consider both facial attributes and expressions.
Thus we propose a novel approach called TEx-Face(TExt & Expression-to-Face) that addresses these challenges by dividing the task into three components, i.e., 3D GAN Inversion, Conditional Style Code Diffusion, and 3D Face Decoding. For 3D GAN inversion, we introduce two methods, which aim to enhance the representation of style codes and alleviate 3D inconsistencies. Furthermore, we design a style code denoiser to incorporate multiple conditions into the style code and propose a data augmentation strategy to address the issue of insufficient paired visual-language data. Extensive experiments conducted on FFHQ, CelebA-HQ, and CelebA-Dialog demonstrate the promising performance of our TEx-Face in achieving the efficient and controllable generation of photorealistic 3D faces. The code will be publicly available.

Controllable 3D Face Generation with Conditional Style Code Diffusion

Deepfake technology has given rise to a spectrum of novel and compelling applications. Unfortunately, the widespread proliferation of high-fidelity fake videos has led to pervasive confusion and deception, shattering our faith that seeing is believing. One aspect that has been overlooked so far is that current deepfake detection approaches may easily fall into the trap of overfitting, focusing only on forgery clues within one or a few local regions. Moreover, existing works heavily rely on neural networks to extract forgery features, lacking theoretical constraints guaranteeing that sufficient forgery clues are extracted and superfluous features are eliminated. These deficiencies culminate in unsatisfactory accuracy and limited generalizability in real-life scenarios.

In this paper, we try to tackle these challenges through three designs: (1) We present a novel framework to capture broader forgery clues by extracting multiple non-overlapping local representations and fusing them into a global semantic-rich feature. (2) Based on the information bottleneck theory, we derive Local Information Loss to guarantee the orthogonality of local representations while preserving comprehensive task-relevant information. (3) Further, to fuse the local representations and remove task-irrelevant information, we arrive at a Global Information Loss through the theoretical analysis of mutual information. Empirically, our method achieves state-of-the-art performance on five benchmark datasets. Our code is available at https://github.com/QingyuLiu/Exposing-the-Deception, hoping to inspire researchers.

Exposing the Deception: Uncovering More Forgery Clues for Deepfake Detection

Cognitive diagnosis is a crucial task in computer-aided education, aimed at evaluating students' proficiency levels across various knowledge concepts through exercises. Current models, however, primarily rely on students' answered exercises, neglecting the complex and rich information contained in un-interacted exercises. While recent research has attempted to leverage the data within un-interacted exercises linked to interacted knowledge concepts, aiming to address the long-tail issue, these studies fail to fully explore the informative, un-interacted exercises related to broader knowledge concepts. This oversight results in diminished performance when these models are applied to comprehensive datasets. In response to this gap, we present the Collaborative-aware Mixed Exercise Sampling (CMES) framework, which can effectively exploit the information present in un-interacted exercises linked to un-interacted knowledge concepts. Specifically, we introduce a novel universal sampling module where the training samples comprise not merely raw data slices, but enhanced samples generated by combining weight-enhanced attention mixture techniques. Given the necessity of real response labels in cognitive diagnosis, we also propose a ranking-based pseudo feedback module to regulate students' responses on generated exercises. The versatility of the CMES framework bolsters existing models and improves their adaptability. Finally, we demonstrate the effectiveness and interpretability of our framework through comprehensive experiments on real-world datasets.

Enhancing Cognitive Diagnosis Using Un-interacted Exercises: A Collaboration-Aware Mixed Sampling Approach

Variational quantum algorithm (VQA) derives advantages from its error resilience and high flexibility in quantum resource requirements, rendering it broadly applicable in the noisy intermediate-scale quantum era. As the performance of VQA highly relies on the structure of the parameterized quantum circuit, it is worthwhile to propose quantum architecture search (QAS) algorithms to automatically search for high-performance circuits. Nevertheless, existing QAS methods are time-consuming, requiring circuit training to assess circuit performance. This study pioneers training-free QAS by utilizing two training-free proxies to rank quantum circuits, in place of the expensive circuit training employed in conventional QAS. Taking into account the precision and computational overhead of the path-based and expressibility-based proxies, we devise a two-stage progressive training-free QAS (TF-QAS). Initially, directed acyclic graphs (DAGs) are employed for circuit representation, and a zero-cost proxy based on the number of paths in the DAG is designed to filter out a substantial portion of unpromising circuits. Subsequently, an expressibility-based proxy, finely reflecting circuit performance, is employed to identify high-performance circuits from the remaining candidates. These proxies evaluate circuit performance without circuit training, resulting in a remarkable reduction in computational cost compared to current training-based QAS methods. Simulations on three VQE tasks demonstrate that TF-QAS achieves a substantial enhancement of sampling efficiency ranging from 5 to 57 times compared to state-of-the-art QAS, while also being 6 to 17 times faster.

Training-Free Quantum Architecture Search

Crime prediction is a crucial yet challenging task within urban computing, which benefits public safety and resource optimization. Over the years, various models have been proposed, and spatial-temporal hypergraph learning models have recently shown outstanding performances. However, three correlations underlying crime are ignored, thus hindering the performance of previous models. Specifically, there are two spatial correlations and one temporal correlation, i.e., (1) co-occurrence of different types of crimes (type spatial correlation), (2) the closer to the crime center, the more dangerous it is around the neighborhood area (neighbor spatial correlation), and (3) the closer between two timestamps, the more relevant events are (hawkes temporal correlation). To this end, we propose Hawkes-enhanced Spatial-Temporal Hypergraph Contrastive Learning framework (HCL), which mines the aforementioned correlations via two specific strategies. Concretely, contrastive learning strategies are designed for two spatial correlations, and hawkes process modeling is adopted for temporal correlations. Extensive experiments demonstrate the promising capacities of HCL from four aspects, i.e., superiority, transferability, effectiveness, and sensitivity.

Hawkes-Enhanced Spatial-Temporal Hypergraph Contrastive Learning Based on Criminal Correlations

Learning correspondences aims to find correct correspondences (inliers) from the initial correspondence set with an uneven correspondence distribution and a low inlier rate, which can be regarded as graph data. Recent advances usually use graph neural networks (GNNs) to build a single type of graph or simply stack local graphs into the global one to complete the task. But they ignore the complementary relationship between different types of graphs, which can effectively capture potential relationships among sparse correspondences. To address this problem, we propose MGNet to effectively combine multiple complementary graphs. To obtain information integrating implicit and explicit local graphs, we construct local graphs from implicit and explicit aspects and combine them effectively, which is used to build a global graph. Moreover, we propose Graph Soft Degree Attention (GSDA) to make full use of all sparse correspondence information at once in the global graph, which can capture and amplify discriminative features.  Extensive experiments demonstrate that MGNet outperforms state-of-the-art methods in different visual tasks. The code is provided in https://github.com/DAILUANYUAN/MGNet-2024AAAI.

MGNet: Learning Correspondences via Multiple Graphs

Compositional Zero-Shot Learning (CZSL) aims to transfer knowledge from seen state-object pairs to novel unseen pairs. In this process, visual bias caused by the diverse interrelationship of state-object combinations blurs their visual features, hindering the learning of distinguishable class prototypes. Prevailing methods concentrate on disentangling states and objects directly from visual features, disregarding potential enhancements that could arise from a data viewpoint. Experimentally, we unveil the results caused by the above problem closely approximate the long-tailed distribution. As a solution, we transform CZSL into a proximate class imbalance problem. We mathematically deduce the role of class prior within the long-tailed distribution in CZSL. Building upon this insight, we incorporate visual bias caused by compositions into the classifier's training and inference by estimating it as a proximate class prior. This enhancement encourages the classifier to acquire more discernible class prototypes for each composition, thereby achieving more balanced predictions. Experimental results demonstrate that our approach elevates the model's performance to the state-of-the-art level, without introducing additional parameters.

Revealing the Proximate Long-Tail Distribution in Compositional Zero-Shot Learning

Video instance segmentation on mobile devices is an important yet very challenging edge AI problem. It mainly suffers from (1) heavy computation and memory costs for frame-by-frame pixel-level instance perception and (2) complicated heuristics for tracking objects. To address those issues, we present MobileInst, a lightweight and mobile-friendly framework for video instance segmentation on mobile devices. Firstly, MobileInst adopts a mobile vision transformer to extract multi-level semantic features and presents an efficient query-based dual-transformer instance decoder for mask kernels and a semantic-enhanced mask decoder to generate instance segmentation per frame. Secondly, MobileInst exploits simple yet effective kernel reuse and kernel association to track objects for video instance segmentation. Further, we propose temporal query passing to enhance the tracking ability for kernels. We conduct experiments on COCO and YouTube-VIS datasets to demonstrate the superiority of MobileInst and evaluate the inference latency on one single CPU core of Snapdragon@778G Mobile Platform, without other methods of acceleration. On the COCO dataset, MobileInst achieves 31.2 mask AP and 433 ms on the mobile CPU, which reduces the latency by 50% compared to the previous SOTA. For video instance segmentation, MobileInst achieves 35.0 AP on YouTube-VIS 2019 and 30.1 AP on YouTube-VIS 2021. Code will be available to facilitate real-world applications and future research.

LF-ViT: Reducing Spatial Redundancy in Vision Transformer for Efficient Image Recognition

Downloads

Next from AAAI 2024

Diversity-Authenticity Co-constrained Stylization for Federated Domain Generalization in Person Re-identification | VIDEO

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES